Article en préparation pour une soumission dans le journal JAC (Journal of antimicrobial chemotherapy)
L’objectif de ce travail était de définir, pour la trithérapie choisie (emtricitabine (FTC), fumarate de ténofovir disoproxil (TDF), dolutegravir (DTG)), un schéma
thérapeutique permettant d’obtenir une concentration plasmatique efficace, c’est-à-dire
dans la zone d’efficacité thérapeutique.
Nous avons étudié les paramètres pharmacocinétiques induits suite à une administration per os ou par voie intrapéritonéale des trois molécules dans des souris Balb/c, BRGSF-A2 non humanisées ou humanisées pour le système immunitaire. Ces études ont permis de (i) confirmer la nécessité d’une prise journalière des ARV du fait de leur demi-vie courte, (ii) valider la stratégie per os, qui si elle est moins précise en termes de doses journalières reçues, est techniquement plus confortable pour les animaux et l’expérimentateur, (iii) définir les concentrations nécessaires pour chacune des drogues. Pour la voie d’administration choisie (la voie per os), nous avons étudié la diffusion tissulaire de ces drogues et prodrogues. Nous avons ainsi confirmé la meilleure pénétration tissulaire du TDF par rapport au FTC et au DTG. Dans le modèle BRGSF-A2, nous retrouvons également la forte pénétration des ARV dans l’intestin, et la faible pénétration de ces molécules dans le cerveau.
Cette étude, prise en charge par la plateforme de pharmacocinétique et pharmacodynamie de l’unité, était un prérequis indispensable à l’établissement d’un modèle d’infection contrôlée par le VIH-1.
1 Pharmacokinetics and tissue distribution of Tenofovir, Emtricitabine and Dolutegravir in a model of mice humanized for the immune system
Hélène Gouget* and Laura Labarthe*, Thibaut Gelé, Mariam-Sarah Benzemrane, Pauline Le Clavez, Nicolas Legrand, Olivier Lambotte, Christine Bourgeois, Roger Le Grand, Aurélie Barrail-Tran
Corresponding author
Aurélie Barrail-Tran
Department of Clinical Pharmacy
AP-HP. Université Paris-Saclay, Hôpital Bicêtre 78 rue du Général Leclerc
94270 Le Kremlin-Bicêtre, France Phone +33 1 45 21 29 64
2
Abstract
Background: Antiretroviral treatments (ART) efficiently control HIV infection, but do not
achieve viral eradication. Combination of ART and various immunotherapies are currently considered. To do so, preclinical model such as mouse model humanized for the immune system (HIS) may prove useful.
Objectives: Study the Pharmacokinetics and tissue distribution of Tenofovir, Emtricitabine
and Dolutegravir in a mouse strain, used as host in humanisation protocols: the BRGSF-A2 strain.
Methods: TNF, FTC and DTG concentrations were simultaneously measured by a LC-
MS/MS method in plasma and tissue samples (including lungs, kidneys, liver, spleen, pancreas, brain, thymus, lymph nodes, caecum, colon, small intestine and adipose tissue). A non-compartmental analysis was performed to estimate the pharmacokinetic parameters.
Results: Concentrations of the three drugs were in the same range of order using the
intraperitoneal or oral route of administration. The trough concentrations for the 3 drugs were close to those observed in patients at therapeutic doses. Comparing the tissue penetration factor TDF was found to diffuse largely in the digestive tract, the kidney, the liver and the pancreas, less in adipose tissue and lung, and not in brain. FTC showed intermediate tissue penetration factor. DTG was poorly distributed to tissues.
Conclusions: ARV pharmacokinetics and tissue distribution in the BRGSF-A2 mouse model
were equivalent to human data. We observed drug specificity (higher penetration of TDF compared to FTC or DTG), and tissue specificity (higher penetration in colon, intestine, kidney, liver versus low penetration in brain (and to a lesser extent lung and adipose tissue)).
3
Introduction
Control of human immunodeficiency virus (HIV) infection can be achieved with optimum antiretroviral therapy (Molina et al., 2015), but does not provide a disease cure. HIV can continue to replicate and replenish the viral reservoir despite potent antiretroviral therapy and plasma HIV-RNA below the limit of quantification (Lorenzo-Redondo et al., 2016). Thus, achieving optimal tissue distribution and cellular pharmacokinetics of antiretroviral drugs would be a prerequisite to the elimination of the viral reservoir and ultimately a step towards a cure for HIV infection.
Studies to map the distribution of antiretrovirals require access to tissues, fluids and cells. In Human, some of these virological reservoirs are hardly or not accessible. Thus, the use of animal models allows a first approach. However, there is no ideal model that reproduces all aspects of HIV infection in a given animal species. Non-human primates are among those most used because of the origin of HIV and the phylogenetic proximity of the simian immunodeficiency virus (SIV) (Worobey et al., 2010). Given the rarity and high cost of these animals, alternative models have been developed. Mouse models humanized for the immune system (HIS) have contributed to a better understanding of the specificity of human immune responses to pathogens. Humanized mice now constitute a translational bridge between fundamental research and clinical applications (Akkina, 2013; Shultz et al., 2012). Developing HIS models enabling to study either the natural course of HIV infection, or controlled HIV infection by antiretrovirals is an important step to evaluate the potential benefit of ARV and immunotherapies, or to dissect the mechanism required for the induction of post-ART control. The Balb/c Rag2KO IL2rgcKO Sirp𝛼NOD Flk2KO HLA-A2HHD (BRGSF-A2) mouse
reconstituted with human CD34+ cord blood cells have been validated as a promising
preclinical model of the mechanisms regulating the expression of various immune checkpoint inhibitors and senescence markers (Labarthe et al., 2020). Considering these results, the aim
4
of this study was to determine doses to reach exposure close to that observed in patients at therapeutic dose, pharmacokinetics parameters and most importantly, tissue disposition for three antiretrovirals in this murine model. Tenofovir (TNF) disoproxil fumarate (TDF), emtricitabine (FTC) and dolutegravir (DTG) were selected as part of the preferred first-line treatment option worldwide (Department of Health and Human Services, 2019; European AIDS Clinical Society, 2019; Saag et al., 2018; World Health Organization, 2019).
5
Methods
Animals
Two different strains of mice were used: wild type Balb/c and the Balb/c Rag2tm1Fwa
IL2rgtm1Cgn SirpaNOD Flk2tm1Irl HLA-A2HHD (BRGSF-A2), a strain with a Balb/c background
bearing multiple mutations generated to allow high human xenograft acceptance (DiSanto et al., 1995; Labarthe et al., 2020; Legrand et al., 2011; Li et al., 2016; Mackarehtschian et al., 1995). Only female animals (that are more prone to human xenograft) were used in this study. Animals from 8 to 18 weeks of age were considered because 18 weeks is the standard time to generate the humanized immune system in the mouse. Animals were weighed before and following ART treatment. Experiments performed on BRGSF-A2 mice and as control on Balb/c animals were approved by the local animal care and use committee (Ethics Committee, CEA Life Science Division (CETEA-CEA DSV IdF CEEA 044), Paris, France) and authorized by the French Ministry of Education and Research (reference: APAFIS #9405 & #9382).
Antiretroviral drugs formulation
The combinatorial antiretroviral therapy (cART) consisted of a three-drug regimen with TDF, FTC and DTG. Two different routes of administration were compared, intraperitoneal and oral. Solution for intraperitoneal (IP) route was prepared by dissolving TDF, FTC and DTG in sterile water using Kleptose HPB parenteral grade (Roquette Pharma, France) as previously described (Del Prete et al., 2016) in order to enhance the solubility and the stability of the molecules. Treatment for oral route was prepared as the IP formulation and then dissolved in a sweetened water gel suspension, MediDrop sucralose (ClearH2O, Westbrook, ME, USA).
6
MediDrop sucralose is an oral formulation that is advertised to assist in the delivery of medications via water bottles.
Antiretroviral drug plasma pharmacokinetic studies
A single IP injection of TDF (5.1 mg/kg), FTC (40.0 mg/kg) and DTG (2.5 mg/kg) was administered to 5 Balb/c mice, 5 BRGSF-A2 mice and 5 human immune system (HIS)- BRGSF-A2 mice. An oral formulation of TDF (47.7 mg/kg), FTC (188.0 mg/kg) and DTG (12.6 mg/kg) was administered during 24 hours via water bottles to 23 Balb/c mice. Blood samples (200 µL) were collected from the anterior facial vein at 1, 2, 4, 6 and 24 hours after the injection or after the end of oral treatment. For the IP injection study, one mouse of each group was sampled at each time. For the oral administration, four mice were sampled at T = 0 h and T = 4 h and 5 mice was sampled at the other time points. Blood samples were centrifuged at 4000xg during 10 minutes at + 4°C and plasma samples were collected and stored at – 80°C until analysis.
Antiretroviral drug tissue pharmacokinetic studies
An oral formulation which combined TDF (47.7 mg/kg), FTC (188.0 mg/kg) and DTG (12.6 mg/kg) was administered during 7 days to 25 Balb/c mice via water bottles. Five mice were euthanized at each time point, at 0, 2, 4, 6 and 24 hours after the end of the treatment. Blood samples were collected at the first time point and tissues including lungs, kidneys, liver, spleen, pancreas, brain, thymus, lymph nodes, caecum, colon, small intestine and adipose tissue were collected at each time point. Blood samples were centrifuged at 4000xg during 10 minutes at + 4°C and plasma samples were collected and stored at – 80°C until analysis. Due to the high proportion found in the feces, especially for DTG (96.7%) (Moss et al., 2015), tissue samples of the gastrointestinal tract (caecum, colon and small intestine) were rinsed
7
with Hank’s Balanced Salt Solution (HBBS) and 2 % of fetal bovine serum. Tissues were snap-frozen and stored at – 80°C until analysis. A tissue penetration factor (TPF) was used to estimate and compare antiretroviral drug penetration. TPF is the ratio of the tissue concentration to the plasma concentration after the tissue concentrations were converted to nanograms per milliliter (to convert volume to mass, tissue density was assumed to be 1.06 g/mL) (Asmuth et al., 2017).
Drug assays
TDF, FTC and DTG concentrations from plasma samples were measured simultaneously by a LC-MS/MS method as previously described (Gouget et al., 2020). After thawing, tissues were diluted in acetonitrile:water (70:30), homogenized with a tissue homogenizer (Minilys®, Bertin Technologies, France) and analyzed by the same method adapted to tissue.
Pharmacokinetic analysis
A non-compartmental analysis using Phoenix WinNonlin 8.1 (Certara, Princeton, NJ, USA)
was performed to estimate the pharmacokinetic parameters. Maximal concentration (Cmax),
trough concentration at steady state (Ct), last measurable concentration (Clast) for single-dose
pharmacokinetics and the time required to reach Cmax (Tmax) are the observed parameters. The
half-life (t½) was estimated on the log-linear terminal part of the phase of decrease in
concentrations according to the formula t½= 0.693/λz where λz is the slope of decrease of the
concentrations. The areas under the curve (AUC) were estimated by the linear up log down trapezoidal method until the last time point (AUC0→t) and up to infinity by extrapolation
(AUC0→∞) using the formula AUC0-t + (Ct/λz). The mean residence time (MRT) was
determined as AUMC/AUC, where AUMC is the area under the moment curve from the time of dosing to the last measurable concentration.
8
Results
Antiretroviral drug plasma pharmacokinetics
Pharmacokinetics parameters and concentrations versus time plots are shown in Table 1 and Figure 1, respectively. Concentrations of the three drugs were in the same range of order in the 3 types of mice and whatever the routes of administration: intraperitoneal or oral. By oral route using antiretroviral treatment via water bottles, 24 hours after the last intake, mean ± standard deviation tenofovir, emtricitabine and dolutegravir concentrations were 25 ± 0 ng /mL, 62 ± 68 ng/mL and 995 ± 509 ng/mL, respectively.
Antiretroviral drugs tissue pharmacokinetics and distribution
All pharmacokinetics parameters in tissue are shown in Table 2. Tenofovir, emtricitabine and dolutegravir concentrations versus time for all tissues are presented in Figure 2. The tissue penetration factors are shown in Figure 3. They ranged from 6.4 to 17459%, 6.5% to 385% and 0.9 to 23.4% for tenofovir, emtricitabine and dolutegravir, respectively. Tenofovir was found to diffuse largely in some tissues as the digestive tract (colon, small intestine, and caecum), the kidney, the liver and the pancreas. Except for adipose tissue, lung and brain, emtricitabine has a good penetration in studied tissues with a TPF up to 100%. Dolutegravir is poorly distributed to tissues. The three drugs have a weak penetration factor in the brain, it could be explaining by the blood-brain-barrier.
9
Discussion
In recent years, acquired immune deficiency syndrome (AIDS) research has become more focused on studies of human immunodeficiency virus (HIV) eradication or functional cure. Mice model have the main advantage to allow for a number of experimental interventions that cannot be conducted in humans. Here in these studies, we attempted to develop a suitable mice model to evaluate antiretroviral drug diffusion. The murine model Balb/c Rag2tm1Fwa IL2rgtm1Cgn SirpaNOD Flk2tm1Irl HLA-A2HHD (BRGSF-A2) was firstly
validated in terms of antiretroviral formulation, dose and route of administration by characterizing the pharmacokinetics of tenofovir disoproxil fumarate, emtricitabine and dolutegravir and comparing the pharmacokinetic parameters according to the route of administration (intraperitoneal versus oral). As the BRGSF-A2 model was similar to a less expensive murine model Balb/c, thereafter we used it to determine antiretroviral tissue distribution. However, because humanisation required irradiation that may affect tissue permeability and penetration, we also performed analyses in mice humanized for the immune system to confirm the results obtained in the Balbc/C and non-humanized BRGSF-A2 mice. Firstly, we compare three different murine models to determine if we observe the same pharmacokinetic profiles and parameters of dolutegravir, emtricitabine and tenofovir after IP administration. Secondly, we used two different ways of administration for the same mouse model. As no previous intraperitoneal or oral pharmacokinetics mice study with our doses was currently reported, we compare the observed concentrations to those observed at therapeutic doses in patients.
Concerning the tenofovir, for all three models and for the two ways of administration, the main pharmacokinetic parameters are in the same range of magnitude. Compare to Human parameters, in our study, the maximal concentrations are higher (median = 2064 ng/mL versus mean = 240 ng/mL in Human (Barditch-Crovo et al., 2001)) and the elimination half-life is
10
smaller (median = 5.7 h versus mean = 11.9 h (Barditch-Crovo et al., 2001)). However, the total drug exposure represented by the AUC is in the same range of magnitude in both species (median = 6012 ng.h/mL versus mean = 3020 ng.h/mL (Barditch-Crovo et al., 2001)). Concerning the emtricitabine, the main pharmacokinetic parameters are also in the same range of magnitude for all three models. Compare to Human parameters, maximal concentrations are also higher (median: 73,899 ng/mL versus 1,820 ng/mL (Blum et al., 2007)). Conversely, trough concentrations are slightly lower (median: 16 ng/mL versus 61 ng/mL (Blum et al., 2007)). These two differences could be partly explained by the difference in half-lives (median: 3.0 h versus 10.1 h (Blum et al., 2007)).
Concerning the dolutegravir, the main pharmacokinetics parameters are also in the same range of magnitude for all three models with greater inter-species variability. Compare to Human parameters, we observed quite the same pharmacokinetics behaviour as emtricitabine with higher maximal concentrations (median: 62,072 ng/mL versus 4,560 ng/mL (Min et al., 2010)) but trough concentrations in the same order of magnitude (median: 2,913 ng/mL versus 1,060 ng/mL (Min et al., 2010)). This also could be explain by the difference in half-lives (median: 7.2 h versus 14.2 h (Min et al., 2010)). However, antiretrovirals can be classified as time- dependent drugs, and, therefore trough concentrations should be above expected levels, namely PA-IC90 (64 ng/mL) (Podany et al., 2020).
The intraperitoneal route has the advantage of knowing with great precision the quantity of antiretrovirals injected. However, it requires handling the mice every day that may cause stress and an increase in complications (peritonitis, intra-intestinal injection). Sterility and pH are also important and constraining parameters of the intraperitoneal route. On the other hand, the oral route has the advantage of requiring no manipulation, nor any physicochemical adjustment. The main inconvenient is the lack of precision in the amount ingested. The use of the oral route reduced the maximal concentrations without modifying the AUC, namely the exposure and without reducing trough concentrations. All main parameters
11
for all three antiretrovirals were consistent and in the same range of magnitude whatever the way of administration. The trough concentrations for the 3 drugs were close to those observed in patients at therapeutic doses. Therefore, the oral route and the antiretroviral doses were validated and were used for tissue distribution studies in Balb/c mice. Few studies had described antiretroviral tissue distributions. Mostly, studies are focused on some tissues, namely lymphoid tissues (Dyavar et al., 2019) or specific tissues like brain (Nicol et al., 2019) or genital tract and rectal tissue (Greener et al., 2013). To our knowledge, we describe in this study the distribution of three major antiretrovirals in all tissues for the first time.
Concerning the tenofovir, trough concentrations and/or tissue penetration factor are quite consistent with already published data in Humans. Very close concentrations were observed in lymph nodes (126 ng/g versus around 100 ng/g (Burgunder et al., 2019)). In the small intestine, some studies focused on more precise and specific parts of this organ, namely on duodenum or ileum. Due to our small model, concentrations are representative of the entire organ and are quite in the same order of magnitude (5,047 (range: 2,864-6635) ng/g versus 9,802 (range: 2,655-17,281) ng/g in ileum (Thompson et al., 2019) or versus 46,100 ng/g in duodenum (Asmuth et al., 2017)). Comparing the tissue penetration factor, data are even closer (13,161% (range:410%-31,333%) versus 8,700% (range: 3,500%-29,700%) (Thompson et al., 2019)). For similar reason, our study focuses on colon rather than the rectum, which is too small in murine model. Nevertheless, concentrations are also in the same order of magnitude (6,994 (range: 3,936-9,518) ng/g versus 1,700 (range: 400-6,700) ng/g (Asmuth et al., 2017) or 1,736 (range: 1,385-3,145) ng/g (Thompson et al., 2019) or 1,877 ng/g (Patterson et al., 2011)). In addition, the tissue penetration factor has a close human value (17,459% (range: 9,062%-28,608%) versus 4,100% (range: 600%-5,200%) (Thompson et al., 2019)). Conversely, in the brain, Nicol M. and colleagues reported higher tenofovir concentrations (range: 161-2,644 ng/g versus 5 (range: 4-7) ng/g (Nicol et al., 2019). This higher tenofovir brain concentration may be explained by much higher tenofovir plasma and
12
tenofovir cerebrospinal fluid concentrations in this study compare to other already published data (Ma et al., 2019). Thus, tissue penetration factors are closer and in the same order of magnitude (6% (range: 3%-11%) in our study versus 36% (range: 14%-124%)).
Concerning the emtricitabine, we also have similar trough concentrations compare to already human published data. In lymph nodes, close concentrations are observed (175 ng/g versus around 10 ng/g (Burgunder et al., 2019)). Closer concentrations are observed in colon compare to rectum human studies (258 (range: 82-419) ng/g versus 1,000 (range: 300-4,600) ng/g (Asmuth et al., 2017) or 600 (range: 146-1,306) ng/g (Thompson et al., 2019) or 124 ng/g (Patterson et al., 2011)). In addition, the tissue penetration factor has a close human value (385% (range: 200%-644%) versus 300% (range: 100%-900%) (Thompson et al., 2019)). At last, small intestine trough concentrations are in the same order of magnitude (421 (range: <LOQ-1,116) ng/g versus 1,200 (range: 400-2,300) ng/g (Asmuth et al., 2017) or 1,019 (range: 389-2,762) ng/g (Thompson et al., 2019)). Thus, tissue penetration factors are also close (295% (range: 90%-632%) versus 700% (range: 200%-1200%) (Thompson et al., 2019)).
Concerning the dolutegravir, very few human data let us compare our model. Nevertheless, Greener B. and colleagues reported dolutegravir pharmacokinetics parameters in colorectal tissue (Greener et al., 2013). Half-life is close to our murine model concentrations (6.8 h versus 8.4 h). Maximal and trough concentrations are also close (Cmax: 996 ng/g versus 418
ng/g; Ct: 104 ng/g versus 139 ng/g). Greener B. and colleagues also reported colorectal
exposition via the AUC0-t and our murine model has a close exposition to dolutegravir
compare to Humans (11,525 ng.h/g versus 7,596 ng.h/g).
By globally observing the distribution of the three antiretrovirals in all the tissues studied, we could first classify these three antiretrovirals according to their diffusion capacity. Tenofovir is the one that diffuses the most in all tissues, then emtricitabine and finally dolutegravir for which all tissue concentrations are below plasma concentrations. These differences can be
13
explained by the physico-chemical and pharmacokinetic differences of these antiretrovirals. For example, tenofovir and emtricitabine have lower molecular weights and are less bound to plasma proteins (Custodio et al., 2016; FDA Approved Drug Products, 2003). On the contrary, dolutegravir is highly bound to plasma protein binding (99%) and substrate of efflux transporters (ABCB1, ABCG2) (Reese et al., 2013). It could explain that less than 25% of plasma dolutegravir is able to penetrate whatever the tissue explored.
There was also an important difference between the different tissues. Nevertheless, there was a certain consistency between all these tissues, namely that the three antiretrovirals studied had a better diffusion in quite the same tissues. The most perfused tissues and metabolism sites (colon, intestine, kidney, liver) were the tissues where diffusion was the most important for the three antiretrovirals. It was only in the lungs among the highly perfused tissues that the diffusion was not as good or even low since it was the only tissue with the brain for which the tissue penetration factor is lower than 100% for all three antiretrovirals. Tissue concentrations have to be compared with 90% inhibitory concentrations (IC90) of the drugs (Flynn et al., 2011; Podany et al., 2020; Thompson et al., 2019). For dolutegravir, only colon and lung concentrations are above PA-IC90 (64 ng/mL). For emtricitabine and tenofovir, all tissue concentrations, brain excluded for emtricitabine, are above the IC90 (51 ng/mL and 2.98